Re: [Idnet] Applying AI into network management//FW: [nmrg] 45th NMRG meeting: Call for Contributions

[Albert Cabellos <[e-mail]>]

Hi all

Below you can find a short report about our work with a
Deep-Reinforcement Learning agent that achieves near-optimal routing
configurations in one single-step (once trained), automatically and
without prior knowledge about the network.

https://arxiv.org/pdf/1709.07080.pdf

We are in the process of scaling-up the experiments.

Kind regards

Albert

On Wed, Sep 20, 2017 at 3:00 PM, Albert Cabellos
<[e-mail]> wrote:
> Hi all
>
> We have been working on using a deep-reinforcement learning agent to
> automatically achieve optimal routing configuration. We demonstrate the
> efficiency of the agent by means of simulations, we will soon (2 weeks
> aprox) make the results public.
>
> The results might be interesting for the IDNET community at large since, at
> the best of our knowledge, this is the first use of deep-reinforcement
> learning for route optimization.
>
> In addition to this, the agent uses a reward function that must be set by
> the operator. This function describes in mathematical terms the desired
> state of the network, for instance to load-balance traffic among the links.
> The agent then aims to maximize the reward function.
>
> This function actually represents the policy set by the orchestrator. In my
> honest opinion there is an interesting discussion on how to express such
> functions in terms of management policy, this might be relevant for the NMRG
> community.
>
> Albert
>
>
> On Tue, Sep 19, 2017 at 4:34 PM, Ciavaglia, Laurent (Nokia - FR/Nozay)
> <[e-mail]> wrote:
>>
>> Dear Yansen, all,
>>
>> We (NMRG chairs) will coordinate with Sheng/IDNET for defining the agenda.
>> Please send your proposal to either lists.
>>
>> Thanks, Laurent.
>>
>>
>> -----Original Message-----
>> From: IDNET [mailto:[e-mail]] On Behalf Of yanshen
>> Sent: Tuesday, September 19, 2017 8:48 AM
>> To: Sheng Jiang <[e-mail]>; [e-mail]
>> Subject: Re: [Idnet] Applying AI into network management//FW: [nmrg] 45th
>> NMRG meeting: Call for Contributions
>>
>> Hi Sheng,
>>
>> I would like to have a short presentation about the Use case of Traffic
>> Prediction/QoS Model.
>>
>> My question is how to "register"? I directly send Email to the NMRG chair
>> or we have a pre-registration in IDNet ?
>>
>> I attach the brief summary of use cases in the end. Hope it helpful.
>>
>> Yansen
>>
>>
>> ==========================================
>> 1. Gap and Requirement Analysis
>>     1.1 Network Management requirement
>>     1.2 TBD
>> 2. Use Cases
>>     2.1 Traffic Prediction
>>                 Proposed by: [e-mail]
>>                 Track:
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00131.html
>>                 Abstract: Collect the history traffic data and external
>> data which may influence the traffic. Predict the traffic in
>> short/long/specific term. Avoid the congestion or risk in previously.
>>
>>     2.2 QoS Management
>>                 Proposed by: [e-mail]
>>                 Track:
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00131.html
>>                 Abstract: Use multiple paths to distribute the traffic
>> flows. Adjust the percentages. Avoid congestion and ensure QoS.
>>
>>     2.3 Application (and/or DDoS) detection
>>                 Proposed by: [e-mail]
>>                 Track:
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00133.html
>>                 Abstract: Detect the application (or attack) from network
>> packets (HTTPS or plain) Collect the history traffic data and identify a
>> service or attack (ex: Skype, Viber, DDoS attack etc.)
>>
>>         2.4 QoE Management
>>                 Proposed by: [e-mail]
>>                 Track:
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00137.html
>>                 Abstract: Collect low-level metrics (SNR, latency, jitter,
>> losses, etc) and measure QoE. Then use ML to understand what is the relation
>> between satisfactory QoE and the low-level metrics. As an example learn that
>> when delay>N then QoE is degraded, but when M<delay<N then QoE is
>> satisfactory for the customers (please note that QoE cannot be measured
>> directly over your network). This is useful to understand how the network
>> must be operated to provide satisfactory QoE.
>>
>>         2.5 (Encrypted) Traffic Classification
>>                 Proposed by: [e-mail]; [e-mail]
>>                 Track: [Jerome]
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00141.html ; [Min-Suk
>> Kim] https://www.ietf.org/mail-archive/web/idnet/current/msg00153.html
>>                 Abstract:
>>                         [Jerome] collect flow-level traffic metrics such
>> as protocol information but also meta metrics such as distribution of packet
>> sizes, inter-arrival times... Then use such information to label the traffic
>> with the underlying application assuming that the granularity of
>> classification may vary (type of application, exact application name,
>> version...)
>>                         [Min-Suk Kim]continuously collect packet data,
>> then applying learning process for traffic classification with generating
>> application using deep learning models such as CNN (convolutional neural
>> network) and RNN (recurrent neural network). Data-set to apply into the
>> models are generated by precessing with features of information from flow in
>> packet data.
>>
>>         2.6 Anomaly Detection
>>                 Proposed by: [e-mail]
>>                 Track:
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00186.html
>>                 Abstract:
>>                         [Jerome] collect flow-level traffic metrics such
>> as protocol information but also meta metrics such as distribution of packet
>> sizes, inter-arrival times... Then use such information to label the traffic
>> with the underlying application assuming that the granularity of
>> classification may vary (type of application, exact application name,
>> version...)
>>                         [Min-Suk Kim]continuously collect packet data,
>> then applying learning process for traffic classification with generating
>> application using deep learning models such as CNN (convolutional neural
>> network) and RNN (recurrent neural network). Data-set to apply into the
>> models are generated by precessing with features of information from flow in
>> packet data.
>>
>> 3. Data Focus
>>     3.1 Data attribute
>>     3.2 Data format
>>     3.3 TBD
>>
>> 4. Support Technologies
>>     4.1 Benchmarking Framework
>>                 Proposed by: [e-mail]
>>                 Track:
>> https://www.ietf.org/mail-archive/web/idnet/current/msg00146.html
>>                 Abstract: A proper benchmarking framework comprises a set
>> of reference procedures, methods, and models that can (or better *must*) be
>> followed to assess the quality of an AI mechanism proposed to be applied to
>> the network management/control area. Moreover, and much more specific to the
>> IDNET topics, is the inclusion, dependency, or just the general relation of
>> a standard format enforced to the data that is used (input) and produced
>> (output) by the framework, so a kind of "data market" can arise without
>> requiring to transform the data. The initial scope of input/output data
>> would be the datasets, but also the new knowledge items that are stated as a
>> result of applying the benchmarking procedures defined by the framework,
>> which can be collected together to build a database of benchmark results, or
>> just contrasted with other existing entries in the database to know the
>> position of the solution just evaluated. This increases the usefulness of
>> IDNET.
>>
>>     4.2 TBD
>>
>> =========================================
>>
>> > -----Original Message-----
>> > From: IDNET [mailto:[e-mail]] On Behalf Of Sheng Jiang
>> > Sent: Wednesday, September 13, 2017 10:20 PM
>> > To: [e-mail]
>> > Subject: [Idnet] Applying AI into network management//FW: [nmrg] 45th
>> > NMRG
>> > meeting: Call for Contributions
>> >
>> > Hi, IDNet,
>> >
>> > After coordinating with NMRG chairs, a Call for Contributions message
>> > (see
>> > below) has been sent by them to the NMRG mailing list regarding to the
>> > topic of applying AI into network management. This is in line with our
>> > earlier discussion to have a session on this in NMRG, Singapore. You
>> > could send email to volunteer for presentations in either NMRG or
>> > IDNet mailing list (I will bridge to NMRG chairs in the IDNet case) or
>> > cross post.
>> >
>> > Looking forward for your contributions and good discussion in Singapore.
>> >
>> > Best regards,
>> >
>> > Sheng
>> >
>> > -----Original Message-----
>> > From: nmrg [mailto:[e-mail]] On Behalf Of Lisandro
>> > Zambenedetti Granville
>> > Sent: Wednesday, September 13, 2017 9:57 PM
>> > To: [e-mail]
>> > Subject: [nmrg] 45th NMRG meeting: Call for Contributions
>> >
>> > Call for Contributions
>> > 45th NMRG meeting at IETF 100
>> >
>> > In the next IETF100/Singapore we will be organizing the 45th NMRG
>> > meeting.
>> > We would like to center the upcoming meeting around the use of
>> > artificial intelligence (AI) for network management, including related
>> > topics as diverse as machine-learning and intelligent-defined networks,
>> > for example.
>> >
>> > AI for network management is not a new topic, as can be easily
>> > observed in the literature produced by the network management community
>> > already years ago.
>> > On the other hand, AI has matured a lot, finding applications is several
>> > areas.
>> > People interested in the subject also formed communities that can
>> > contribute too. As such, revisiting AI for network management is not
>> > only appropriate but also timely.
>> >
>> > In this Call for Contributions we would like to receive proposals of
>> > presentations/discussions for the upcoming 45th NMRG meeting. That
>> > includes, for example:
>> >
>> > - Use cases where AI could/should be used in network management
>> > - Real-life experiments, results, and findings on AI for network
>> > management
>> > - Disruptive and/or new management paradigms based on AI
>> > - Potential standard requirements for applying AI for network
>> > management
>> > - Both preliminary and mature approaches
>> >
>> > Please contribute and feel free to distribute this call to other
>> > mailing lists whose members you believe would be interested and could
>> > contribute too.
>> >
>> > Best regards, Lisandro and Laurent.
>> > _______________________________________________
>> > nmrg mailing list
>> > [e-mail]
>> > https://www.irtf.org/mailman/listinfo/nmrg
>> > _______________________________________________
>> > IDNET mailing list
>> > [e-mail]
>> > https://www.ietf.org/mailman/listinfo/idnet
>>
>> _______________________________________________
>> IDNET mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/idnet
>>
>> _______________________________________________
>> IDNET mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/idnet
>
>